1. Virtual Machines
A Teaching Case in the Systems
Administration Course
Iwan Tjhin
Whitireia New Zealand
iwan.tjhin@whitireia.ac.nz
2. Introduction
» The Importance
» Common Problems
» computer lab setup
» no dedicated machines for students
» Common Solutions
» local virtualisation
» centralised prefabricated virtualisation
3. Limitations
Potentially deny students the opportunity to gain
the first-hand experience they need to move
successfully into a systems administration role.
4. Objectives
» A flexible learning environment
» All VMs are centrally managed and run
» Without the restriction of existing IT policy
Minimise common solutions' limitations
5. Objectives
» Performance
» at least 8 VMs for each student, run concurrently
» High availability
» continuous access after class time from remote locations
» Monitoring progress
» monitor and check on students’ works progress remotely
» Interference-free
» no interference to Faculty's network
» Submission
» simplifed submission logistics
6. Methodology
» July 2012 semester
» 17 students enrolled, given access and used the system
» 15 weeks course
» 13 class weeks (1-10 and 13-15)
» 2 mid-term study-break weeks (week 11 to 12)
» Sample selection
» Successfully submitted assignment
» Did not withdraw from the course
» Attended timetabled class regularly
» Only 8 students matched the criteria
7. Methodology
» The Infrastructure
» High Performance Computing Center (HPCC)
» 338 IBM blades
» > 2.0 THz of CPU power
» > 2.6 TB of RAM capacity
» 28 blades of IBM BladeCenter HS20 32-bit model
» Summary of software installed:
» Blade 1 to 26 : ESXi host
» Blade 27 : Base OS + vCenter server
» Blade 28 : Base OS + RDP server + vSphere client + recording tool
» vCenter server was kept in a separate blade
» Load balancing reason
8. Methodology
» Virtual Machines and Operating Environment
» VM files stored on local blade disk drives
» 8 VMs per student (136 VMs in total)
» 3 VMs in a test environment
» for practice exercises and learning
» 5 VMs in a production environment
» for working on marked assignment tasks
11. Methodology
» Beginning of Semester
» Briefing and giving access accounts
» Gave a set of lab class exercises (on test environment VMs)
» Gave a set of assignment tasks (on production environment VMs)
» End of Semester
» Disabled access to system
» Marking only on VMs in the production environments
13. • High Availability Result
– Over total access
• Outside class time = 58%
• During weekends = 18%
• At nights = 17%
– During vs. outside class time
• Average access
– up from 10.62 to 14.53
• Median access
– up from 10.50 to 13.00
14. Limitations and Future Work
» Logon recording did not include the duration of each logon
» No record or indication of each logon's duration
» Might correlate with work effort
» Data sampling is small due to the nature of class size
» Study could be repeated on different and larger groups
» Additional test variables
» For a greater understanding
» Effects on students’ skills set and study experience
15. Conclusion
» Study suggested that with the system:
» Impracticality and expense could be minimised
» Has interference-free network operation environment
» Has high availability, including access outside of class time
» Able to run more VMs concurrently, simulating real-life situations more closely
» Able to monitor students’ work progress and provided timely feedback
» Simplified submission logistics
» Marking without relying on screenshots, portable medias or marking “on the spot” in
class time
16. References
» Ahmad, I., Anderson, J. M., Holler, A. M., Kambo, R., & Makhija, V. (2003). An analysis of disk performance in VMware ESX server virtual machines. Workload Characterization
2003, (pp. 65-76).
» Begnum, K., Koymand, K., Krap, A., & John, S. (2004). Using virtual machines in system administration education. Proceedings of the 4th International System Administration
and Network Engineering Conference. Amsterdam.
» Berqia, A., Diop, A., & Harms, J. (2002). A virtual laboratory for practical exercises. Proceedings of the International Conference on Engineering Education, (pp. 1-5).
Manchester.
» Chaudhary, V., Cha, M., Walters, J. P., Guercio, S., & Gallo, S. (2008). A comparison of virtualization technologies for HPC. 22nd International Conference on Advanced
Information Networking and Applications (AINA) 2008 (pp. 861-868). IEEE.
» Cranitch, G., & Rees, M. J. (2009). Virtualisation: A case study in database administration laboratory work. Proceedings ascilite Auckland 2009, (pp. 168-174).
» Gaspar, A., Langevin, S., & Armitage, W. D. (2007). Virtualization technologies in the undergraduate IT curriculum. IT Pro, pp. 36-43.
» Grammer, K., Stolerman, J., & Yi, B. (2011). Introduction of virtualization in the teaching of operating systems for CS undergraduate program. Journal of Computing Sciences in
Colleges, 26(6), 44-50.
» Jin, K., & Miller, E. L. (2009). The effectiveness of deduplication on virtual machine disk images. Proceedings of SYSTOR 2009: The Israeli Experimental Systems Conference
(pp. 9-18). ACM.
» Kfir, R. E. (2001). Virtual laboratories in education. Proceedings of the 1st International Conference on Computer Graphics, Virtual Reality and Visualisation (pp. 27-31). New
York: ACM New York.
» Kumar, R., & Singh, G. (2012). Learning computer networking using virtualization tools. Computing and Information Technology Research and Education New Zealand, (pp.
52-55). Christchurch.
» Mishchenko, D. (2011). Install ESXi 3.5 to an IDE drive. Retrieved June 19, 2012, from http://vm-help.com/esx/esx3i/ESXi_install_to_IDE_drive/ESXi_install_to_IDE_drive.php
» Stackpole, B., Koppe, J., Haskell, T., Guay, L., & Pan, Y. (2008). Decentralized virtualization in systems administration education. Proceedings of the 9th ACM SIGITE conference
on Information technology education (pp. 249-254). ACM.
» Steele, A. (2010). Using Google Docs for the early identification of 'at risk' students. Proceedings of Conference of Computing and Information Technology Research and
Education New Zealand, (pp. 261-266). Dunedin.
» Stockman, M. (2003). Creating remotely accessible "virtual networks" on a single PC to teach computer networking and operating systems. Proceeding of the 4th Conference
on Information Technology Education (pp. 67-71). ACM.
» Vollrath, A., & Jenkins, S. (2004). Using virtual machines for teaching system administration. Journal of Computing Sciences in Colleges, 20(2), 287-292.
» White, J., & Pilbeam, A. (2010). A survey of virtualization technologies with performance testing. arXiv preprint arXiv:1010.3233.
» Yang, L. (2007). Teaching system and network administration using Virtual PC. Journal of Computing Sciences in Colleges, 23(2), 137-142.
» Younge, A. J., Henschel, R., Brown, J. T., Laszewski, G. v., Qiu, J., & Fox, G. C. (2011). Analysis of virtualization technologies for high performance computing environments. 4th
International Conference on Cloud Computing (pp. 9-16). IEEE.
Importance
Real-life work experience
First-hand experience and skills set
Successful Transition
Systems administration role
Cloud computing and virtualisation driven environment
Local Virtualisation
computer availability
performance of lab computers
storage
lab computer network interference
monitoring students' progress
submission and marking
Centralised Prefabricated Virtualisation
inflexible, prefabricated and preconfigured
tight IT security policy applied
restricted access
far from real-life situation
VMware ESXi 3.5 (“ESXi host”)
VMware vCenter Server 4.0 (“vCenter server”)
Microsoft Server 2008 (“base OS”)
Remote Desktop Protocol server (“RDP server”)
VMware vSphere Client 4.0 (“vSphere client”)
Logon recording tool ("recording tool")
Connection Design and Access Method
A single-point access management configuration
The RDP server as central contact point
Accessed through RDP thin-clients
Remote sessions directly from LAN or SSH over the Internet
vSphere client launched from within the remote session
VMs operations directly through this vSphere client
Connect to client OSs via RDP or web browser
Unified Access Permission
2-stage logon process
RDP server
vCenter server
When connecting from the Internet
Additional SSH tunnel logon
Unified unique logon accounts for each individual student
Created using Microsoft Powershell scripts
Logon Recording
Logon recording tool installed
Using FrontRunnerTek UserLog v1.0
Supports command line script execution
Logon script using .bat file
Placed at the RDP server’s logon policy
Ran whenever a student logged-on
Network Isolation Design
Multiple VLAN tags, each tag unique to an individual student’s VM environment
Each student’s set of VMs were grouped to two different VLANs
1 for test environment
1 for production environment
Resulted in a large overall operating environment with many VLANs
No. of VLANs = No. of students in class x No. of VLAN for each student
Access to the external network routed through separate NIC
Broadcast packets were limited and restricted to own individual VLAN environments
Performance Result
All 136 VMs ran successfully
Every student ran concurrently all 8 VMs
Able to work on assignment tasks using the VMs as intended
High Availability Result
VMs remained running outside of class time
Access to system and resume working from remote locations
Monitoring Progress Result
Teaching staff gained access to all VMs
Monitor students’ work progress remotely
Accessed, checked, and given feedback weekly
No waiting until next class time
Class time freed up for teaching
Interference-free Result
No associated interruption was detected or reported
DHCP requests and broadcasts did not exit the isolated network operating environments
Submission and Marking
No screenshots
No electronic media submission
No huge files to transfer
Disabled access
Marking done at time and pace suitable to the teaching staff
This indicates that the solution provides high availability by enabling continuous access for students within and outside class sessions, including from remote locations.
The outcome was that, by being available to students at all time, the system was able to accommodate both fluctuating and high demand levels.
This finding indicates that students were quick to take advantage of the greater learning opportunities provided by the system.